Aktív témák

• #4570 Gorneck legenda

  2007-01-20 23:15:07 #4570

  Összes hozzászólása itt Válaszok az összes hozzászólására itt Válaszok erre a hozzászólásra

  Privát üzenet küldése

  

  Gorneck

  legenda

  XS:
  ''I think I have found the main reason for our cold boot problem. I hooked up a dmm to the read point for vmch.
  
  I started out with plans this afternoon of being vdroop, vmch, and vdimm. Vdroop so I could get a more reasonable vcore at idle. Vdimm since mine was 0.1v lower than what was set. Plus, I figured if I made starting vdimm at 2.1v, it may eliminate some boot issues. Results show vdimm boost didn't help boot issues though, but it's nice to have +0.2v have it at 2.3v.
  
  Vmch. Here's the interesting guy. It's been brought up many times before, and people have done actual reads based on different fsbs and all settings normal, +0.1, +0.2, and +0.3. I should have done this as a baseline for comparison but I didn't think to at the time. I ran out of qtips for the time being, but I need to erase the pencil for vmch anyway.
  
  But here's what I found. I recall setting it so that base vmch was 1.35v. Then I proceeded to try and do 450fsb. I tried +0.2v setting as my preference, since I read I think a vrzone article saying they stopped seeing cold boot issues after keeping vmch at 1.55+. However, this +0.2v resulted in a setting somewhere around 1.7v. What the... Maybe the graphite heated up and become less resistive. I'll check that in a bit. I try normal and +0.3v and both send me to 1.8v vmch. Yikes. +0.1 sets me very close to my original target of 1.55 though so I figured I'd just try to keep it there.
  
  One thing I noticed was that each time the system went for a reboot, I saw a first vmch setting of 1.45-1.49v regardless of what vmch is set to. Then it pulls it up to what the setting was. However, I haven't been able to boot 450fsb ever since I did the mod. Got it to post maybe once or twice, with insane vmch iirc. But even then it didn't boot successfully into windows.
  
  I'll try to get rid of the graphite, and see if I can determine any characteristics of it at stock.
  
  Perhaps if there was a way to clamp vmch at 1.55 - 1.6v, everything would be peachy? The problem is that I'm not sure how to approach this. Since with however much graphite I put on, it looks like the effective baseline vmch is being set to 1.45-1.49, which rarely do I see applied. The problem is that ''normal'' goes to the +0.3v equiv at the fsb's we want to be focusing on. And the problem with that comes in that there is no +0.0. The best compromise would be +0.1v. If things do work out properly, this would mean it'd always turn on at 1.55v, and once it gets ready to post it pulls it up to 1.65v. With adequate cooling, 1.65 should be ok. However, this still has the problems of normal/+0.3 sending vmch to nearly 1.9v (seems the +0.3v setting is slightly more than a 0.3v jump). On recovery posts, I believe it always forces max vmch, so there should be frequent shocks for us tweakers.
  
  I'd imagine that clamping vmch would require soldering, beyond the pencilmods I did.
  
  
  edit: Oh yeah, I nuked my windows again. Agh. I shorted out the read point for vmch with one of the legs, powering off the system. On boot up again, bios settings are retained, except time gets reset. I didn't realize this, and proceeded to try and boot windows. I was greeted with ''trial period over'' again. Grrr, back to safemode to backup and format once I get it up I guess
  
  update: bleh didn't help with instability. guess the board was returned for a reason. but woohoo, got my rev2 working again.''

Aktív témák